To examine the relationship between intramyocardial pH during global ischemic arrest and subsequent functional and biochemical recovery, 40 canine hearts were subjected to 4 hours of arrest at 10 degrees C. Four groups, each containing 10 hearts, were differentiated by the oxygen concentration of a hyperkalemic crystalloid cardioplegic solution (CCS), which was infused every 20 minutes. In group 1 the CCS was equilibrated at 4 degrees C with nitrogen to remove oxygen. In group 2 the CCS was aerated at 4 degrees C. In group 3 the CCS was treated to achieve an oxygen tension (PO2) similar to group 2 but with a reduced nitrogen content to prevent bubble formation, which is theoretically possible during reperfusion ("myocardial bends"). In group 4 the CCS was fully oxygenated at 4 degrees C. The resulting PO2 of CCS measured at 10 degrees C was less than 20, 170, 170, and 750 mm Hg in groups 1, 2, 3, and 4, respectively. Left ventricular function (LVF) was assessed from function curves at constant mean aortic pressure and heart rate. Functional recovery, expressed as a percentage of prearrest LVF, was 38.1% +/- 10.7% in group 1 and 84.0% +/- 8.1% in group 4 (p less than 0.008). Functional recovery was 64.9% +/- 5.5% and 69.1% +/- 7.0% in groups 2 and 3, which had similar PO2. Differences in recovery between groups 2 and 3 and group 1 approached statistical significance (p less than 0.05, NS). The mean-integrated intramyocardial pH during arrest was higher (p less than 0.003) in group 4 (7.14 +/- 0.05) than in group 1 (6.84 +/- 0.06) or group 2 (6.86 +/- 0.07). The minimum intramyocardial pH during arrest was higher in group 4 than in any other group (p less than 0.002). Myocardial adenosine triphosphate concentration at the end of arrest, expressed as a percentage of its prearrest value, was highest in group 4 (75.9% +/- 8.1%) and lowest in group 1 (54.3% +/- 5.7%), a difference approaching statistical significance (p less than 0.05, NS). These data suggest that the measurement of intramyocardial pH is a useful on-line indicator of the adequacy of preservation during hypothermic arrest and that excess nitrogen in aerated CCS had little or no effect on recovery. The data confirm the hypothesis that oxygenation of CCS is associated with good myocardial preservation, which may be attributed to the provision of oxygen for the support of aerobic metabolism during arrest.